Padma Shri A. S. Kiran Kumar of Former Chairman, ISRO

Lunar exploration missions of India have enhanced scientific understanding of the Moon’s surface composition, thermal characteristics, seismic activity, plasma environment, and near-surface conditions. They have also contributed to the development of critical capabilities in spacecraft navigation, propulsion, deep-space communication, autonomous landing, rover operations, and in situ scientific experimentation.
The knowledge, experience, and technological capabilities gained through the Chandrayaan programme are expected to influence India’s future missions to the Moon and beyond. Theyprovide a foundation for advanced lunar research, sample-return missions, international scientific collaboration, exploration of lunar resources, and future human lunar exploration. India’s journey to the Moon is therefore not only a significant national achievement but also an important contribution to humanity’s continuing exploration of space and an inspiration for future generations of scientists and engineers.

Address:Orissa, India

Dr. Radhakant Padhi of HAG Professor & Prof. Satish Dhawan Chair Department of Aerospace Engineering , Indian Institute of Science (IISc), Bengaluru, India

High-precision autonomous soft-landing is increasingly becoming an essential requirement for various interplanetary missions, and also for stage recovery of launch vehicles for reusability. From the guidance control perspective, the problem can be broadly classified into two categories: (i) landing on atmosphere-less planets and satellites (such as Moon) and (ii) landing on planets and satellites with atmosphere (such as Earth and Mars). Optimal guidance plays a very important role in both, as it can generate optimal trajectories satisfying system dynamics, boundary conditions, applicable constraints and yet ensuring a meaningful optimality (such as acceleration/jerk/fuel optimality). This talk will give an overview of both aspects of this critical technology.
First part of this talk will remain focused on the Chandrayaan mission of India. It will include some of the lessons learnt from the unsuccessful attempt of Chandrayaan-2, followed by the improvements carried out subsequently that led to the success of the Chandrayaan-3 mission. Focus will be on the guidance system improvement, including the contributions made by the speaker and his research group. Next, the launch vehicle stage recovery problem will be discussed, including the challenges it offers. To tackle this problem, a strong pitch will be made in favour of the nonlinear optimal control theory based next-generation computational guidance, and why India is ready for this challenge.

Address:Orissa, India

Shri Prof. Sankar Kumar Pal of ANRF Prime Minister Professor, GoI, at IIIT Bhubaneswar

The talk first describes the components of granular computing (GrC) and information granules, significance of rough sets in granular computing, and the relevance of generalized rough sets as a stronger paradigm for uncertainty modeling. The problem of unsupervised video tracking in ambiguous situations is considered as an example application of granular mining. Roles of different spatio-color and spatio-temporal granules as well as the lower and upper approximations are explained.
The second part deals with the significance of deep learning (DL) in object recognition and tracking. After explaining the relevance of granulation in CNN (convolutional neural nets) for reducing the computation time, a deep architecture, named G-RCNN (granulated region proposal neural nets) is presented. This is an improved version of the Fast RCNN and Faster RCNN for extracting RoIs (regions of interest) by incorporating spatio-temporal granulation in a deep AlexNet architecture. G-RCNN accepts videos directly as input. Spatio-temporal granules enable extraction of less but more representative RoI-pixels of object regions and thereby make the G-RCNN superior in real-time detection speed and accuracy.
As a case study, the merits of G-RCNN in terms of speed, localization and detection accuracy are demonstrated on real-time traffic scenarios in a Tata Steel Plant for anomaly detection towards safety analytics. Here the problem of “fall” and “no-fall” detection is dealt with as well as the anomaly between “collision” and “near-miss”. Significance of Z-numbers as a measure of certainty (trust) in the occurrence of collision and near-miss is illustrated. This concept is unique.
The talk concludes by mentioning the challenging issues and future directions of research including the evolution of discipline since 1960s and some concerns for beginners.